Threshold protective levels of serum IgG to Shigella lipopolysaccharide: re-analysis of Shigella vaccine trials data

OBJECTIVES: Establishing a correlate of protection is essential for the development and licensure of Shigella vaccines. We examined potential threshold levels of serum IgG to Shigella lipopolysaccharide (LPS) that could predict protection against shigellosis. METHODS: We performed new analyses of serologic and vaccine efficacy (VE) data from two randomized vaccine-controlled trials of the Shigella sonnei–Pseudomonas aeruginosa recombinant exoprotein A (rEPA) conjugate conducted in young adults and children aged 1–4 years in Israel. Adults received either S. sonnei–rEPA (n = 183) or control vaccines (n = 277). Children received the S. sonnei–rEPA conjugate (n = 1384) or S. flexneri 2a–rEPA conjugate (n = 1315). VE against culture-proven shigellosis was determined. Sera were tested for IgG anti–S. sonnei LPS antibodies. We assessed the association of various levels of IgG anti–S. sonnei LPS antibodies with S. sonnei shigellosis risk using logistic regression models and the reverse cumulative distribution of IgG levels. RESULTS: Among adults, four vaccinees and 23 controls developed S. sonnei shigellosis; the VE was 74% (95% CI, 28–100%). A threshold of ≥1:1600 IgG anti–S. sonnei LPS titre was associated with a reduced risk of S. sonnei shigellosis and a predicted VE of 73.6% (95% CI, 65–80%). The IgG anti–S. sonnei LPS correlated with serum bactericidal titres. In children, a population-based level of 4.5 ELISA Units (EU) corresponding to 1:1072 titre, predicted VE of 63%, versus 71% observed VE in children aged 3–4 years. The predicted VE in children aged 2–4 years was 49%, consistent with the 52% observed VE. CONCLUSION: Serum IgG anti–S. sonnei LPS threshold levels can predict the degree of VE and can be used for the evaluation of new vaccine candidates.